#Desactivar warnings
import warnings
warnings.filterwarnings('ignore')
warnings.simplefilter(action='ignore', category=Warning)

#Carga de datos inicial
import pandas as pd
datos = pd.read_excel("hprice3.xlsx")

#Nuevas variables
datos["Bathrooms"] = datos["baths"]
datos["Area (sqft)"] = datos["area"]
datos["Log Price ($)"] = datos["log_price"]
datos["Neighborhood"] = datos["neighborhood"].astype(str)
datos["Year"] = datos["year"].astype(str)
datos["neighborhood_str"] = datos["neighborhood"].astype(str)

#Datos en formato largo
tabla = pd.crosstab(datos[datos["neighborhood"] != 0]["neighborhood"], datos['baths'])
datos_largo = tabla.melt()
datos_largo["neighborhood"] = list(range(1, 7))*4
datos_largo["Neighborhood"] = list(range(1, 7))*4
datos_largo["Baths"] = datos_largo["baths"].astype(str)
datos_largo["baths"] = datos_largo["baths"].astype(str)
datos_largo["Frequency"] = datos_largo["value"]

#Paleta de colores
colores = ["paleturquoise", "lightskyblue", "royalblue", "blue"]
colores_bp = ['steelblue', 'salmon']

import matplotlib.pyplot as plt

# Gráfico de dispersión
plt.figure()
for neighborhood in sorted(datos["neighborhood"].unique()):
    neigh_data = datos[datos["neighborhood"] == neighborhood]
    plt.scatter(x=neigh_data["area"],
                y=neigh_data["log_price"],
                label=neighborhood,
                alpha=0.95,
                marker=".",
                cmap='Set2'
                )

plt.title("Area of the house and selling price by neighborhood")
plt.xlabel('Area (sqft)')
plt.ylabel('Log Price ($)')
plt.legend(title='Neighborhood', loc=4)
plt.show()

# Boxplot
plt.figure()
bp = plt.boxplot(x=[datos[datos["year"] == 1978]["log_price"], datos[datos["year"] == 1981]["log_price"]],
                 labels=["1978", "1981"],
                 patch_artist=True
                 )

plt.title('Selling prices by year')
plt.xlabel('Year')
plt.ylabel('Log Price ($)')


for patch, color in zip(bp['boxes'], colores_bp):
    patch.set_facecolor(color)
for median in bp['medians']:
    median.set(color='black',
               linewidth=2)
plt.show()

# Gráfico de barras
plt.figure()
bott = 0
for baths in sorted(datos["baths"].unique()):
    plt.bar(sorted(datos[datos["neighborhood"] != 0]["neighborhood"].unique()),
            pd.crosstab(datos[datos["neighborhood"] != 0]["neighborhood"], datos['baths'])[baths],
            bottom=bott,
            label=f'{baths}',
            color=colores[baths-1],
            edgecolor='black')
    bott = bott + pd.crosstab(datos[datos["neighborhood"] != 0]["neighborhood"], datos['baths'])[baths]
plt.title('Number of bathrooms in the house by neighborhood')
plt.xlabel('Neighborhood')
plt.legend(title='Bathrooms', loc=1)
plt.show()

import seaborn as sns

# Gráficos de dispersión
plt.figure()
sns.scatterplot(data=datos,
                x="area",
                y="log_price",
                hue="neighborhood",
                palette="deep"
                )
plt.title("Area of the house and selling price by neighborhood")
plt.xlabel('Area (sqft)')
plt.ylabel('Log Price ($)')
plt.legend(title='Neighborhood', loc=4)
plt.show()

# Boxplot
plt.figure()
sns.boxplot(data=datos, x="year", y="log_price",
            palette=colores_bp,
            width=0.25)
plt.title('Selling prices by year')
plt.xlabel('Year')
plt.ylabel('Log Price ($)')
plt.show()

# Gráfico de barras
plt.figure()
sns.histplot(
    data=datos[datos["neighborhood"] != 0],
    x="neighborhood",
    hue="Bathrooms",
    multiple="stack",
    discrete=True,
    shrink=0.8,
    palette=colores
)
plt.title('Number of bathrooms in the house by neighborhood')
plt.xlabel('Neighborhood')
plt.show()

from plotnine import (
    ggplot,
    aes,
    geom_point,
    geom_boxplot,
    geom_bar,
    theme_bw,
    theme_gray,
    theme_classic,
    labs,
    scale_fill_brewer
)

# Gráfico de dispersión
p = (
    ggplot(datos, aes("area", "log_price", color="neighborhood_str"))
    + geom_point()
    + labs(x='Area (sqft)',
           y='Log Price ($)',
           color="Neighborhood",
           title="Area of the house and selling price by neighborhood")
    + theme_bw()
)
print(p)
# Boxplot
p = (
    ggplot(datos)
    + geom_boxplot(aes(x="factor(year)",
                       y="log_price",
                       fill="factor(year)"
                       ))
    + labs(x="Year",
           y='Log Price ($)',
           title="Selling prices by year")
    + theme_gray()
)
print(p)

# Gráfico de barras
p = (
    ggplot(datos[datos["neighborhood"] != 0])
    + geom_bar(aes(x="neighborhood", fill="factor(baths)"), color="black")
    + labs(x="Neighborhood",
           fill="Bathrooms",
           title="Number of bathrooms in the house by neighborhood")
    + scale_fill_brewer()
    + theme_classic()
)
print(p)

import plotly.express as px

# Gráfico de dispersión
fig = px.scatter(datos.sort_values(by="neighborhood"),
                 x="area",
                 y="log_price",
                 color="neighborhood_str",
                 color_discrete_sequence=px.colors.qualitative.Set1,
                 symbol="neighborhood_str",
                 hover_name="price",
                 width=800,
                 height=800,
                 labels={
    "area": "Area (sqft)",
    "log_price": "Log Price ($)",
    "neighborhood_str": "Neighborhood"
}
)
fig.update_layout(title="Area of the house and selling price by neighborhood")
fig.update_traces(
    hovertemplate='Area: %{x} sqft<br>Price: $%{hovertext} <br><extra></extra>'
)
fig.show()

# Boxplot
fig = px.box(datos,
             x="year",
             y="log_price",
             color="year",
             color_discrete_sequence=colores_bp,
             width=800,
             height=800,
             labels={
                 "year": "Year",
                 "log_price": "Log Price ($)"
             }
             )
fig.update_layout(title="Selling prices by year")
fig.show()

# Gráfico de barras
fig = px.bar(datos_largo,
             x="neighborhood",
             y="value",
             color="baths",
             color_discrete_sequence=colores,
             width=800,
             height=800,
             labels={
                 "neighborhood": "Neighborhood",
                 "value": "Count",
                 "baths": "Bathrooms"
             }
             )

fig.update_layout(title="Number of bathrooms in the house by neighborhood")
fig.show()

import bokeh.palettes
from bokeh.transform import factor_cmap
from bokeh.plotting import figure, output_file, show
from bokeh.io import output_notebook

# Gráfico de dispersión
colores_cmap = factor_cmap('neighborhood_str', palette=bokeh.palettes.Set3[7],
                      factors=sorted(datos.neighborhood_str.unique()))

graph = figure(title="Area of the house and selling price by neighborhood")
graph.scatter("area", "log_price", source=datos,
              size=12.5, 
              fill_alpha=0.6,
              color="black", 
              fill_color=colores_cmap,
              legend_group='neighborhood_str'
              )
graph.xaxis.axis_label = 'Area (sqft)'
graph.yaxis.axis_label = 'Log Price ($)'
graph.legend.title = 'Neighborhood'
graph.legend.location = "bottom_right"
output_notebook()
show(graph)

# Gráfico de barras
neighborhood = ['1', '2', '3', '4', '5', '6']
baths = ['1', '2', '3', '4']
datos_barras = {'neighborhood': neighborhood,
                '1': tabla[1],
                '2': tabla[2],
                '3': tabla[3],
                '4': tabla[4]
                }

graph = figure(title="Number of bathrooms in the house by neighborhood",
               x_range=neighborhood, 
               tooltips="Frequency: @$name")
graph.vbar_stack(baths, x='neighborhood',
                 width=0.9,
                 color=colores,
                 source=datos_barras,
                 legend_label=baths
                 )

graph.xaxis.axis_label = 'Neighborhood'
graph.yaxis.axis_label = 'Frequency'
output_notebook()
show(graph)

import altair as alt
alt.renderers.enable('html')

# Gráfico de dispersión
chart = alt.Chart(datos, title="Area of the house and selling price by neighborhood").mark_circle(size=60).encode(
    alt.X('Area (sqft)').scale(zero=False),
    alt.Y('Log Price ($)').scale(zero=False),
    color='Neighborhood',
    tooltip=["Area (sqft)", "Log Price ($)", "Neighborhood"]
).interactive()
chart.display()

# Boxplot
chart = alt.Chart(datos, title="Selling prices by year").mark_boxplot(size=50).encode(
    alt.X('Year').scale(zero=False),
    alt.Y('Log Price ($)').scale(zero=False),
    alt.Color("Year").scale(scheme="set1").legend(None),
).properties(
    width=250
)
chart.display()

# Gráfico de barras
chart = alt.Chart(datos_largo, title="Number of bathrooms in the house by neighborhood").mark_bar(size=30).encode(
    x='Neighborhood',
    y='Frequency',
    order=alt.Order('Neighborhood', sort='ascending'),
    color=alt.Color('Baths').scale(scheme="blues"),
    tooltip=["Neighborhood", "Baths", "Frequency"]
)
chart.display()

import pygal
from pygal.style import Style


# Gráfico de dispersión
chart1 = pygal.XY(stroke=False)
chart1.title = 'Area of the house and selling price by neighborhood'
chart1.x_title = 'Area (sqft)'
chart1.y_title = 'Log Price ($)'

for x in range(0, 7):
    serie = list(zip(
        datos[datos["neighborhood"] == x]["area"].values,
        datos[datos["neighborhood"] == x]["log_price"].values
    ))
    chart1.add(str(x), serie)
    pass

display(chart1)

# Boxplot
chart2 = pygal.Box(box_mode="tukey")
chart2.title = 'Selling prices by year'
chart2.x_title = 'Year'
chart2.y_title = 'Price ($)'
chart2.add("1978", datos[datos["year"] == 1978]["price"].values)
chart2.add("1981", datos[datos["year"] == 1981]["price"].values)

display(chart2)

# Gráfico de barras
estilo = Style(colors=(colores))

chart3 = pygal.StackedBar(style=estilo)
chart3.title = 'Number of bathrooms in the house by neighborhood'
chart3.x_title = 'Neighborhood'
chart3.x_labels = list(range(1, 7))

for x in range(1, 5):
    chart3.add(str(x), tabla[x].values)
    pass

display(chart3)

# Configurar formato HTML
from IPython.display import display, HTML
base_html =  u"""
    <!DOCTYPE html>
    <html>
        <head>
            <script type="text/javascript" src="http://kozea.github.com/pygal.js/javascripts/svg.jquery.js"></script>
            <script type="text/javascript" src="http://kozea.github.com/pygal.js/javascripts/pygal-tooltips.js"></script>
        </head>
        <body><figure>{rendered_chart}</figure></body>
    </html>
"""

# Mostrar los gráficos
display(HTML(base_html.format(rendered_chart=chart1.render(is_unicode=True))))
display(HTML(base_html.format(rendered_chart=chart2.render(is_unicode=True))))
display(HTML(base_html.format(rendered_chart=chart3.render(is_unicode=True))))

Visualización de datos con Python¶

Tabla de contenidos¶

Configuración inicial ¶

Matplotlib ¶

Seaborn ¶

Plotnine ¶

Plotly ¶

Bokeh ¶

Altair ¶

Pygal ¶